0
\$\begingroup\$

Can someone explain how this circuit is working? How is the victim node voltage increasing through the leakage current?

Here is the image for more context:

Crosstalk image and text explanation

(Source: Team VLSI - Crosstalk Noise and Crosstalk Delay – Effects of Crosstalk)

\$\endgroup\$
9
  • \$\begingroup\$ A question to think about: What is the impedance of a capacitor, as seen by high-frequency transients? \$\endgroup\$
    – nanofarad
    Commented Oct 10, 2022 at 1:55
  • \$\begingroup\$ The impedance is very low, right? \$\endgroup\$
    – KEE97
    Commented Oct 10, 2022 at 1:57
  • \$\begingroup\$ @KEE97 - Hi, This is clearly a follow-on from your previous question and since you already got answers there which seem to answer the questions currently asked above, please edit this question and add more detail of exactly how the previous answers weren't enough. Otherwise you may waste time (both yours and the people replying) by again getting similar answers to those you got on that previous question. Thanks. \$\endgroup\$
    – SamGibson
    Commented Oct 10, 2022 at 2:01
  • \$\begingroup\$ I would genuinely just prefer an explanation of the highlighted line in the image, that’s all! \$\endgroup\$
    – KEE97
    Commented Oct 10, 2022 at 2:13
  • \$\begingroup\$ @KEE97 - Re: "I would genuinely just prefer an explanation of the highlighted line in the image" To me, answers to your previous question do explain that highlighted line, which is why I asked you to add more detail to avoid wasted time. We'll see if someone wants to try explaining in a different way to before. Thanks. \$\endgroup\$
    – SamGibson
    Commented Oct 10, 2022 at 2:20

1 Answer 1

2
\$\begingroup\$

It might be easier for you to picture it if you instead interpret the impedances in the system as resistances.

The impedance of a capacitor is very low at high frequencies. Think about what the transient on the aggressor line looks like in the frequency domain - the faster the transient, the higher frequency components it must have.

The impedance of the driver for each line (aggressor and victim) is not zero. The traces on the PCB have some parasitic resistance, inductance, and reactance.

If you mentally picture these impedances as resistors, in a kind of instantaneous snapshot of the circuit's operation, you can get a feel for why a voltage is induced in the victim line:

schematic

simulate this circuit – Schematic created using CircuitLab

The actual numbers here are arbitrarily chosen. All that matters is the overall behaviour.

The aggressor line has an impedance of 1Ω from its driver. The same applies to the victim line. The capacitor has an instantaneous impedance of 10Ω during the transient.

This forms a divider circuit. We can calculate the induced voltage in the victim line (i.e. the voltage at B) as:

$$V_{B} = \frac {Z_2} {Z_1 + Z_C + Z_2} \times V_1 = \frac {1\Omega} {1\Omega + 10\Omega + 1\Omega} \times 1V = 83mV$$

Assuming equal impedance for both drivers during the transient, we can also show that the aggressor line suffers a voltage dip equal to the induced voltage in the victim line:

$$V_{A} = \frac {Z_1 + Z_C} {Z_1 + Z_C + Z_2} \times V_1 = \frac {1\Omega + 10\Omega} {1\Omega + 10\Omega + 1\Omega} \times 1V = 0.917V$$

Now think about the situation at low frequencies or DC. Capacitors have high impedance at low frequency, so \$Z_C\$ will be very high. Let's arbitrarily pick 100kΩ as an example:

$$V_{B} = \frac {Z_2} {Z_1 + Z_C + Z_2} \times V_1 = \frac {1\Omega} {1\Omega + 100000\Omega + 1\Omega} \times 1V \approx 10\mu V$$

There is still a very small amount of induced voltage in the victim line, but it is far less due to the high impedance of the capacitor.

\$\endgroup\$
4
  • 1
    \$\begingroup\$ This makes a lot of sense when I look at it like this and in terms of the formulas and math. I also wanted to think about it in terms of currents and properties of a capacitor and failed to understand, was wondering if you could shed some light! Thank you so much for your response. \$\endgroup\$
    – KEE97
    Commented Oct 10, 2022 at 3:05
  • \$\begingroup\$ There's no magic with the current. A current flows when there's a potential difference between two points. The magnitude of the current that flows is inversely proportional to the impedance. The impedance of a mutual capacitance is dependent on the frequency of the signals along with the physical properties of the two transmission lines and the environment. Ultimately it's about the behaviour of the electric fields around the transmission line; typically you would use a field solver to figure out what the coupling will look like for a given set of transmission lines on a PCB. \$\endgroup\$
    – Polynomial
    Commented Oct 10, 2022 at 3:14
  • \$\begingroup\$ perhaps it might help the OP, in addition to the excellent explanation above, to consider that at the moment when the transistion occurs there is a a short burst of HF signal in the frequency domain which is not there when the signal is in steady state. \$\endgroup\$
    – danmcb
    Commented Oct 10, 2022 at 7:58
  • \$\begingroup\$ @KEE97 Suppose I have a circuit with a resistor, capacitor, another resistor. Let's say 100\$\mu\$F and 100k\$\Omega\$ (quite big so the time constant is long). Then I tap my magic wand on one side of the capacitor and add some voltage to it. In the next microsecond, what happens to the voltage across the capacitor? Which equation calculates the voltage across a capacitor? \$\endgroup\$ Commented Oct 10, 2022 at 10:44

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.